############# Obfuscator/ #############

Obfuscator.py - program to set up and run obfuscation. Grondahl.py - obfuscates using grondahl techniques viper_*.py - VIPER program, WRITTEN BY VIPER AUTHORS full code found: https://github.com/UKPLab/naacl2019-like-humans-visual-attacks

############# ObfuscatorReddit/ #############

Obfuscator.py - program to set up and run obfuscation for reddit data.

############ Detectors/ ############

NULI.py - implementation of NULI offenseval system

vradivchev.py - implementation of vradivchev offenseval 2019 system

MIDAS.py - implementation of MIDAS offenseval 2019 system

Perspective.py - Perspective offense classifier implementation, api key needed

LexiconDetect.py - Lexicon offense classifier.

########## DetectorsReddit/ ##########

NULI.py - implementation of NULI trained on reddit data

MIDAS.py - implementation of MIDAS trained on reddit data

########## Datasets/ ##########

CensoredRedditData_ratio_15.0.tsv - dataset for reddit 1_train_CensoredRedditData_ratio_15.0.tsv - training split for reddit results 1_test_CensoredRedditData_ratio_15.0.tsv - testing split for reddit results

NonOffensiveTweets - Final list of non offensive tweets as deemed by classifiers (Make up evasion dataset) createWordEmbeddingsNoConstraint.py - script to create NEW embeddings trainWordEmbeddingsNoConstraint.py - script to fine tune (FT) glove embedding (NOTE trained glove embeddings too large for github will link another way)

subset_OFF_testset-taska-gold.tsv - subset of offensive examples from test set from offenseval, used in results offenseval-training-v1.tsv - training set from offenseval, used to train models in results